[MemCpyOpt] Support opaque pointers

diff --git a/llvm/lib/IR/IRBuilder.cpp b/llvm/lib/IR/IRBuilder.cpp
index ca65773..3c55b73 100644 (file)
--- a/llvm/lib/IR/IRBuilder.cpp
+++ b/llvm/lib/IR/IRBuilder.cpp
@@ -62,7 +62,7 @@ Type *IRBuilderBase::getCurrentFunctionReturnType() const {
 
 Value *IRBuilderBase::getCastedInt8PtrValue(Value *Ptr) {
   auto *PT = cast<PointerType>(Ptr->getType());
-  if (PT->getElementType()->isIntegerTy(8))
+  if (PT->isOpaqueOrPointeeTypeMatches(getInt8Ty()))
     return Ptr;
 
   // Otherwise, we need to insert a bitcast.

diff --git a/llvm/lib/Transforms/Scalar/MemCpyOptimizer.cpp b/llvm/lib/Transforms/Scalar/MemCpyOptimizer.cpp
index db8d40b..6016779 100644 (file)
--- a/llvm/lib/Transforms/Scalar/MemCpyOptimizer.cpp
+++ b/llvm/lib/Transforms/Scalar/MemCpyOptimizer.cpp
@@ -1215,8 +1215,7 @@ bool MemCpyOptPass::processMemSetMemCpyDependence(MemCpyInst *MemCpy,
   Value *MemsetLen = Builder.CreateSelect(
       Ule, ConstantInt::getNullValue(DestSize->getType()), SizeDiff);
   Instruction *NewMemSet = Builder.CreateMemSet(
-      Builder.CreateGEP(Dest->getType()->getPointerElementType(), Dest,
-                        SrcSize),
+      Builder.CreateGEP(Builder.getInt8Ty(), Dest, SrcSize),
       MemSet->getOperand(1), MemsetLen, MaybeAlign(Align));
 
   if (MSSAU) {

diff --git a/llvm/test/Transforms/MemCpyOpt/memset-memcpy-redundant-memset.ll b/llvm/test/Transforms/MemCpyOpt/memset-memcpy-redundant-memset.ll
index 10ec8b4..9873216 100644 (file)
--- a/llvm/test/Transforms/MemCpyOpt/memset-memcpy-redundant-memset.ll
+++ b/llvm/test/Transforms/MemCpyOpt/memset-memcpy-redundant-memset.ll
@@ -291,10 +291,27 @@ define void @test_must_alias_different_size(i8* noalias %src, i64 %src_size, i8*
   ret void
 }
 
+define void @test_opaque_ptrs(ptr %src, i64 %src_size, ptr noalias %dst, i64 %dst_size, i8 %c) {
+; CHECK-LABEL: @test_opaque_ptrs(
+; CHECK-NEXT:    [[TMP1:%.*]] = icmp ule i64 [[DST_SIZE:%.*]], [[SRC_SIZE:%.*]]
+; CHECK-NEXT:    [[TMP2:%.*]] = sub i64 [[DST_SIZE]], [[SRC_SIZE]]
+; CHECK-NEXT:    [[TMP3:%.*]] = select i1 [[TMP1]], i64 0, i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP4:%.*]] = getelementptr i8, ptr [[DST:%.*]], i64 [[SRC_SIZE]]
+; CHECK-NEXT:    call void @llvm.memset.p0.i64(ptr align 1 [[TMP4]], i8 [[C:%.*]], i64 [[TMP3]], i1 false)
+; CHECK-NEXT:    call void @llvm.memcpy.p0.p0.i64(ptr [[DST]], ptr [[SRC:%.*]], i64 [[SRC_SIZE]], i1 false)
+; CHECK-NEXT:    ret void
+;
+  call void @llvm.memset.p0.i64(ptr %dst, i8 %c, i64 %dst_size, i1 false)
+  call void @llvm.memcpy.p0.p0.i64(ptr %dst, ptr %src, i64 %src_size, i1 false)
+  ret void
+}
+
 declare void @llvm.memset.p0i8.i64(i8* nocapture, i8, i64, i1)
 declare void @llvm.memcpy.p0i8.p0i8.i64(i8* nocapture, i8* nocapture readonly, i64, i1)
 declare void @llvm.memset.p0i8.i32(i8* nocapture, i8, i32, i1)
 declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture readonly, i32, i1)
 declare void @llvm.memset.p0i8.i128(i8* nocapture, i8, i128, i1)
 declare void @llvm.memcpy.p0i8.p0i8.i128(i8* nocapture, i8* nocapture readonly, i128, i1)
+declare void @llvm.memset.p0.i64(ptr nocapture, i8, i64, i1)
+declare void @llvm.memcpy.p0.p0.i64(ptr nocapture, ptr nocapture readonly, i64, i1)
 declare void @call()